[TUHS] First Unix-like OSes not derived from AT&T code?

Bakul Shah bakul at iitbombay.org
Tue May 3 02:16:40 AEST 2022


On May 2, 2022, at 8:43 AM, Clem Cole <clemc at ccc.com> wrote:
> 
> Thoth Thucks....   [Kelly Booth gave me one of these tee's years go].
> 
> Mike Malcolm did not try to clone UNIX - for one thing, it was in B [which Steve Johnson has spread the gospel of same on his sabbatical).  It was not until the Thoth rewrite that became QNX that they tried to ensure all of the Unix behaviors and APIs.  Mike was certainly had an influence by UNIX and IIRC his thesis and the Thoth papers reference/compare it.

IIRC Gordon Bell and Dan Dodge worked with Thoth as students but QNX is not derived from it. I ran across QNX at a contract job in 1986 or so[1]. Back then it fit in 8KB. IIRC the original few versions were mostly written in assembly language or had substantial portions in assembly while most of Thoth was written in C[2]. The original QNX was basically a message passing microkernel. Unix APIs came in much later.

[1] I had to debug some obscure problem where the QNX was running on a text to speech board plugged in a PC and was connected to an IBM 370 system. The TTS system was used to allow a salesperson to place an order or something. It failed randomly. In the end it turned out to be an "undocumented" h/w bug in 286 (Intel knew about it but they denied when we asked!). I caught it in the act using a logic analyzer! Anyway, I had to get pretty familiar with QNX then.

[2] In 1981 AMD tried to get into computer business via AMC (American Micro Computers). They  used Thoth! I interviewed there but in the end joined Fortune Systems.


More information about the TUHS mailing list